Classification and Regression Trees, Cart TM - A user manual for identifying indicators of vulnerability to famine and chronic food insecurity + disk (IFPRI, 1999, 56 p.)
 (introduction...) Preface Acknowledgments 1. Introduction 2. Overview of CART 3. Basic Principles of Cart Methodology 4. Regression Trees: An Overview 5. CART Software and Program Codes 6. Refining CART Analyses 7. Conclusions Appendix 1: Condensed Examples of Classification-Tree Output (full output on diskette) Appendix 2: A Condensed Example of Regression-Tree Output (full output on diskette) References Back Cover

### Appendix 2: A Condensed Example of Regression-Tree Output (full output on diskette)

 Node 1 was split on variable NDVIMNMXA case goes left if variable NDVIMNMX £ 0.335000Improvement = 95.212097 C.T. = 0.439885E+0.05 Figure Node Cases Average Standard Deviation 1 462 10.902165 19.447115 2 174 23.455744 25.835199 -3 288 3.317708 7.119483 Surrogate Split Association Improvement 1 KRMTMNMN s 0.335000 0.931034 88.178185 2 NDVIMXMX s 0.475000 0.827586 84.152443 3 BEGAMXMN s 0.365000 0.793103 62.045887 4 BEGAMNMN s 0.300000 0.793103 68.704895 5 KRMTMXMN s 0.475000 0.793103 79.163147 Competitor Split Improvement 1 KRMTMNMN 0.335000 88.177315 2 NDVIMXMX 0.475000 84.151741 3 KRMTMX 0.435000 81.044876 4 KRMTMXMN 0.475000 79.162216 5 KRMTMN 0.285000 78.868150 Node 2 was split on variable MZSHTTRDA case goes left if variable MZSHTTRD £ 31.389999Improvement = 58.391998 C.T = 0.269771E+0.05 Figure Node Cases Average Standard Deviation 2 174 23.455744 25.835199 -1 65 39.580002 28.492435 -2 109 13.840366 18.272223 Surrogate Split Associations Improvement 1 MZSHTTDV s -0.780000 0.861539 38.192043 2 CERLPPDV s -0.085000 0.492308 26.714172 3 BELGMNDV s -0.855000 0.492308 39.261301 4 BEGAMNDV s -0.620000 0.369231 18.076315 5 KRMTMNDV s -1.150000 0.323077 27.505999 Competitor Split Improvement 1 BELGMNDV -0.905000 43.058678 2 BEGAMNDV -0.055000 40.379753 3 MZSHTTDV -1.185000 39.435616 4 KRMTMXDV -1.905000 37.680218 5 KRMTMNDV -0.900000 32.776260

Note: C.T. stands for Complexity Threshold (the complexity parameter used in tree pruning).